Tech Founder Says He Was ‘Pretty Wrong’ About Lean Teams in the AI Era

As some corporations race to make use of AI to shrink their groups, one startup founder goes in opposition to the grain by planning to rent extra individuals.

Yasser Elsaid, the founding father of customer-service AI startup Chatbase, mentioned the corporate now has 40 staff and plans to double its staff over the subsequent six months.

“I may need been improper about constructing with a lean staff within the age of AI,” Elsaid wrote in a publish on X this month.

Elsaid beforehand embraced smaller groups. In a LinkedIn post in January, he mentioned Chatbase, a San Francisco-based startup that lets corporations construct AI agents for customer support and assist, had stayed at 17 full-time staff for a lot of its first three years, with most hires in engineering.

Nevertheless, in feedback despatched to Enterprise Insider, Elsaid mentioned that he now considers his guess on lean groups “fairly improper.”

“I feel it is a large mistake for founders to stay in a future that does not exist but,” he mentioned.

Elsaid mentioned that extra succesful coding models, together with Anthropic’s newest Claude mannequin Opus 5.5, initially made him suppose Chatbase might get by with fewer individuals.

However Elsaid mentioned that this conclusion solely holds if a startup retains the identical roadmap it had earlier than AI. Fairly than settle for that code had turn out to be a commodity, he determined Chatbase ought to set its sights larger.

“If constructing is simpler, construct more durable issues,” he mentioned. “Most moats in software program come from somebody doing one thing onerous for longer, and being proper. That hasn’t gone away.”

‘Not totally unbiased’

That view runs counter to a broader push towards AI-powered lean teams.

Block reduce practically half its workforce in February, with CEO Jack Dorsey saying AI and “smaller and flatter groups” had been altering how corporations are constructed.

Coinbase, in the meantime, reduce 14% of employees in Could, citing engineers’ ability to ship in days what as soon as took groups weeks; it’s now testing smaller AI-native pods and “one individual groups.”

The identical logic has powered the “tiny groups” startup development: OpenAI CEO Sam Altman has predicted a one-person billion-dollar company, whereas Cursor grew from $1 million to $100 million in annual recurring income in below a yr with fewer than 50 staff, based on non-public market analysis platform Sacra.

However Elsaid mentioned one of the best fashions will not be but unbiased sufficient to exchange individuals. They nonetheless want people to resolve what to construct, give them route, and ensure the outcome works for purchasers, he mentioned.

“Code is basically proper or improper: it does what you need or it does not,” Elsaid mentioned. “However what to construct, the way it ought to work and what the expertise ought to really feel like is what makes your product totally different.”

He estimated that about 90% of Chatbase engineers’ work is now product work: researching, speaking to clients, forming an opinion, and turning that right into a specification an agent can code.

The remaining requires engineers who can perceive the code and ensure totally different AI brokers will not be working in opposition to one another, he mentioned.

Chatbase can be hiring customer-facing employees, significantly in gross sales and buyer success, he mentioned, as a result of enterprise clients nonetheless have to be heard and persuaded.

The fashions are “extraordinarily good and succesful, however not totally unbiased,” he added. “Perhaps that modifications and sooner or later you will not want people within the constructing course of in any respect. However that is not at present.”